mcpbeat Sign in

MAD Synapse · Web & Research MCP Server

answering

MAD Synapse · Web & Research is answering right now. Last checked 11 min ago. It exposes 15 tools.

Read the web (page text, DNS, WHOIS, TLS, uptime) + research (Wikipedia, arXiv, GitHub, packages).

Uptime history 14 days of history · worst day 99%
14 days agonow
100.0%
Uptime 24h
91 of 91 checks
15
Tools
read from the server
550 ms
Response time
average over 24h
open, no key
Access
streamable-http

What changed 1

Every tool that appeared, vanished or quietly changed what it asks for. Recorded since 10 October 2026. No other catalogue keeps this.

10 Oct a tool changed version

Nothing serious here today

Today is the operative word: we check MAD Synapse · Web & Research every 15 minutes and re-read its code on every release. Watch it and you find out the day that stops being true.

Three servers free · no card

Connect this server

Endpoint below is the one we actually reach during checks — not the one copied from a README. Last verified 11 min ago.

run in your terminal
claude mcp add synapse-web --transport http https://agent.maddegen.art/hub/mcp/web
~/Library/Application Support/Claude/claude_desktop_config.json
{
  "mcpServers": {
    "synapse-web": {
      "url": "https://agent.maddegen.art/hub/mcp/web"
    }
  }
}
~/.codex/config.toml
[mcp_servers.synapse-web]
url = "https://agent.maddegen.art/hub/mcp/web"
.cursor/mcp.json
{
  "mcpServers": {
    "synapse-web": {
      "url": "https://agent.maddegen.art/hub/mcp/web"
    }
  }
}
.vscode/mcp.json
{
  "mcpServers": {
    "synapse-web": {
      "url": "https://agent.maddegen.art/hub/mcp/web"
    }
  }
}

Available tools 15

Read directly from the server with tools/list, grouped by what they act on. If a tool disappears, we record the date.

web
web_meta
Everything a page says about itself: title, description, OpenGraph/Twitter card, canonical, favicon, language, RSS/Atom feeds, JSON-LD schema types, headings — and link-preview problems. Use to build link previews, check SEO/social cards, find a site's feeds or structured data. Flags missing og:image, missing description and non-absolute image URLs (common reasons a link shows up bare on X/Telegram). When to use: For title/OpenGraph/canonical; for the article text use web_read. Price: $0.001 per call (10 free/day; after that a payment-required result lists x402 options). Errors: returns isError with a message for invalid input or an upstream failure (not charged).
web_read
Any public URL → clean readable markdown: the article/main content without nav, ads or boilerplate, plus title, author, date, word count and links. PDFs too. Fetches the page (redirects followed, public internet only), extracts the main content with Mozilla Readability and converts it to markdown. PDFs are converted to text. Use max_chars to bound the output. JavaScript-only pages return whatever the server sends (no headless browser). When to use: For a page's content; for its metadata only use web_meta; for uptime/redirects use http_probe. Price: $0.002 per call (10 free/day; after that a payment-required result lists x402 options). Errors: returns isError with a message for invalid input or an upstream failure (not charged).
arxiv
arxiv_search
Search arXiv research papers by keywords, author or category — newest or most relevant — with abstracts, authors, dates and PDF links. arXiv's official API. query supports arXiv syntax (ti:, au:, abs:, cat:cs.AI, AND/OR). Paced to arXiv's one-request-per-3-seconds policy and cached, so repeated queries are instant. Price: $0.001 per call (10 free/day; after that a payment-required result lists x402 options). Errors: returns isError with a message for invalid input or an upstream failure (not charged).
book
book_search
Find books by title, author, subject or ISBN: first published year, editions, ISBNs, page count, ratings, subjects and cover image. Open Library (Internet Archive) catalogue of 30M+ titles. Price: $0.001 per call (10 free/day; after that a payment-required result lists x402 options). Errors: returns isError with a message for invalid input or an upstream failure (not charged).
dns
dns_lookup
All DNS for a domain — A, AAAA, CNAME, MX, NS, TXT, CAA, SOA — plus SPF and DMARC parsed and graded, so an agent knows where a domain points and whether its email can be spoofed. Queried live against public resolvers (Cloudflare, Google, Quad9). Email grade: SPF present and ending in -all/~all, DMARC present with p=quarantine/reject. When to use: For DNS records; for ownership dates use domain_whois; for certificates use tls_check. Price: free. Errors: returns isError with a message for invalid input or an upstream failure (not charged).
doi
doi_lookup
Resolve a DOI — or search a paper title — to full citation metadata: title, authors, journal, year, citations count, publisher, license, and a ready-made APA/BibTeX citation. Crossref's open metadata (150M+ works). Search mode returns the top matches for a title/author query. When to use: For citation metadata of a known paper; to search preprints by topic use arxiv_search. Price: $0.001 per call (10 free/day; after that a payment-required result lists x402 options). Errors: returns isError with a message for invalid input or an upstream failure (not charged).
domain
domain_whois
Who registered a domain and when: registrar, creation/expiry dates, age, days to expiry, status locks, nameservers — or that it looks available to register. Uses RDAP, the registries' official successor to WHOIS (rdap.org bootstrap). A domain with no RDAP record and no DNS is reported as likely available. Young domains (< 30 days) are flagged: most phishing/scam sites are days old. When to use: For registration and expiry; for DNS records use dns_lookup. Price: $0.002 per call (10 free/day; after that a payment-required result lists x402 options). Errors: returns isError with a message for invalid input or an upstream failure (not charged).
email
email_check
Will this email address work? Syntax, domain MX records, disposable/throwaway provider detection, role address (info@, admin@) and free-mail provider — without sending anything. No SMTP probing (that gets servers blacklisted); the verdict is from syntax, live MX lookup and a maintained list of ~100k disposable domains. Price: $0.001 per call (10 free/day; after that a payment-required result lists x402 options). Errors: returns isError with a message for invalid input or an upstream failure (not charged).
github
github_repo
A GitHub repository at a glance: stars, forks, open issues, last commit, latest release, license, languages, contributors, archived — is it alive and trustworthy? From the GitHub REST API. "activity" summarizes days since last push and last release so an agent can judge whether a project is maintained. Price: $0.002 per call (10 free/day; after that a payment-required result lists x402 options). Errors: returns isError with a message for invalid input or an upstream failure (not charged).
hn
hn_search
Search Hacker News stories and comments by keyword, sorted by relevance or date, filtered by time window and minimum points — what developers are saying about anything. Algolia's official HN search. Returns title, url, points, comment count, author, date and the HN discussion link. Price: $0.001 per call (10 free/day; after that a payment-required result lists x402 options). Errors: returns isError with a message for invalid input or an upstream failure (not charged).
http
http_probe
Is a site up and how is it served: status, full redirect chain, timings (DNS/TLS/first byte/total), server, CDN, caching, compression, and a graded security-header check (HSTS, CSP, frame, sniffing…). One real GET from our server. The security grade counts HSTS, Content-Security-Policy, X-Frame-Options/frame-ancestors, X-Content-Type-Options, Referrer-Policy and Permissions-Policy. When to use: For availability, status and timings; for page content use web_read. Price: $0.001 per call (10 free/day; after that a payment-required result lists x402 options). Errors: returns isError with a message for invalid input or an upstream failure (not charged).
package
package_info
Any npm, PyPI or crates.io package: latest version and release date, license, downloads, maintainers, repo, dependency count, deprecation — to pick or vet a dependency. Straight from each registry (plus npm and pypistats download counts). Flags deprecated packages and ones with no release in 2+ years. When to use: For package metadata; for known vulnerabilities in a version use vuln_check (on the MAD Synapse · Wallets & Risk server, https://agent.maddegen.art/hub/mcp/risk). Price: $0.001 per call (10 free/day; after that a payment-required result lists x402 options). Errors: returns isError with a message for invalid input or an upstream failure (not charged).
robots
robots_check
May I crawl this? Checks a URL against the site's robots.txt for any user-agent, lists sitemaps and crawl-delay, and reports which AI crawlers (GPTBot, ClaudeBot, PerplexityBot, Google-Extended…) the site blocks. Standard robots.txt matching (longest match wins, Allow beats Disallow on ties, * and $ wildcards). Missing robots.txt = everything allowed. Price: $0.001 per call (10 free/day; after that a payment-required result lists x402 options). Errors: returns isError with a message for invalid input or an upstream failure (not charged).
tls
tls_check
A site's TLS certificate: issuer, validity, days until expiry, SANs, protocol, cipher, chain and whether browsers trust it. Live TLS handshake to the host. Flags expired, expiring (< 14 days), untrusted or name-mismatched certificates. Price: $0.001 per call (10 free/day; after that a payment-required result lists x402 options). Errors: returns isError with a message for invalid input or an upstream failure (not charged).
wiki
wiki
Wikipedia in any language: search a topic and get the best article's summary, description, thumbnail and link — or the full article text. Search + REST summary from Wikipedia. full=true returns the article as plain text (sections kept), capped by max_chars. Disambiguation pages are flagged with alternatives. When to use: For encyclopedic topics; for papers use arxiv_search or doi_lookup. Price: free. Errors: returns isError with a message for invalid input or an upstream failure (not charged).

Endpoints

URLTransportStateLatencyChecked
https://agent.maddegen.art/hub/mcp/web streamable-http answering 427 ms 11 min ago

Alternatives to MAD Synapse · Web & Research

same job, measured the same way
RepoFold
by repofold

A cited wiki of your GitHub repo: search, read pages, find symbols and ask, with line citations.

answering
Web & Developer Data Tools
by projectworks007

Website tech stacks, domain WHOIS/DNS/SSL, sitemaps, PDF text, GitHub trending and more.

answering
Ctx
by vanshulgoyal101

Turn a GitHub repo or docs site into agent-ready context: pack it or search it, over MCP.

6 tools answering
Code of Paper
by isagelswann

Find GitHub implementations of arXiv research papers — paper lookup, ranked code, search.

205 installs/wk local only
Developer Tools MCP Server
by thenextgennexus

Access GitHub repos, npm/PyPI packages, Stack Overflow, arXiv, and Google Scholar

16 installs/wk local only
Herohunt MCP
by herohunt-ai

People search API - find 1 billion candidates across LinkedIn and GitHub with real-time data.

30 installs/wk local only
Grok Build
by nuruvala

Run the Grok Build CLI as MCP tools: agent runs with live progress, code review, web research.

89 installs/wk local only
Aur MCP
by mrfentmen

Arch User Repository package search from the AUR RPC. No key required.

82 installs/wk local only

MAD Synapse · Web & Research — questions

Answers built from our own checks of this server.

What can MAD Synapse · Web & Research do?
It exposes 15 tools, read directly from the server on our last check. Among them: arxiv_search, book_search, dns_lookup, doi_lookup, domain_whois, email_check and 9 more. The full list with descriptions is on this page — we take it from the server itself via tools/list, not from a README. How MCP servers expose tools in the first place →
Is MAD Synapse · Web & Research working right now?
We send a real MCP handshake every 15 minutes. Over the last 24 hours 91 of 91 checks got a reply (100.0%), average response time 550 ms. The bar chart above shows every period we have measured.
How do I connect MAD Synapse · Web & Research?
Copy the ready config from this page — we generate it for Claude Code, Claude Desktop, Codex, Cursor and VS Code, each with the file path that client actually reads. It is a remote server, so there is nothing to install — the client connects to the address.
Does MAD Synapse · Web & Research need an API key?
No. MAD Synapse · Web & Research completed a full MCP handshake with us as an anonymous client and listed its tools without asking for anything. All 15 of them are readable on this page. This is what we observed, not what the docs claim.
How fast is MAD Synapse · Web & Research?
It answers our handshake in 550 ms on average, which is faster than 22% of all working MCP servers we measure. That is on the slow side — worth knowing if the tool sits inside an interactive loop. The comparison comes from our own checks across the whole registry, every 15 minutes.